The DNA microarray allows for the simultaneous measurement of thousands of genes in one experiment. Microarray-derived gene expression data is highly dimensional, making its analysis difficult due to the well-known curse of dimensionality [16]. Because many genes in these datasets are unrelated to illness, efficient feature selection is necessary to enhance the precision and readability of predictive models. Binary Differential Evolution (BDE) with a newly developed mutation operator has been recently used for biclustering [3]. This work proposes a novel feature selection strategy based on multi-objective optimization to identify a reduced subset of significant genes from breast cancer microarray data. Extensive experiments on a breast cancer dataset demonstrated the effectiveness of the proposed technique in terms of classification accuracy.